Data Analysis Checklist

Pre Data Collection

• Know you Goals (This is a big and vague)
• Divide your goals in different sections (Example, Demographics, Performance, Motivation)
• Create Questions
  • Ask yourself if the question should be in close ended (quantitative) or open ended (qualitative/mixed) question
    • Open Ended (Short and Long Text) - Essay, Follow-Up, Free Response
    • Close Ended (Checkbox, Radio Buttons) - Multiple Choice, Scales, Date, Binary Question
  • Make sure that it standardized (look for reference on different question banks)
  • Done in putting all of the REQUIRED questions (no more, no less)
• Add response validations
• Is question correlated to another question?

Pause Point

Before you actually distribute the data collection tool

• Have you double checked all of the questions?
• Create a test response
• Let others check your questionnaire for suggestions

During Collection

• If Possible, setup a semi-automated cleaning in your spreadsheet
  • Trim whitespace
  • Clean the illegal characters
  • Set the case of string (Proper, Upper, Lower)
  • Replace (RegEx, Find, Left, Right, Search) the unnecessary details
  • Split the responses - Checkbox
  • Concatenate the responses - Scale (Multi Row/Column), or responses that is separated with validation
• Check if there is an error popping out while data collection - Optional but recommended
• Spread it like a fire (More responses, more happiness in visualization (JOKE)) - Quality data first before quantity this is related with sampling

After Collection

• Divide your data into 70:30 randomly (To eliminate bias, do an analysis in the 70% of data and keep the 30% until you finish your analysis. Then do the exact analysis to the 30% and compare it to the 70%.)

For Analysis Steps

• Aggregate the data (Please Study)
• Perform statistical test (HAHAHA) Most of the time Descriptive and Slight Inferential Stats
• Check the correlations/comparisons of data
• Check If the analysis of 70 and 30 matches
  • If yes, proceed to data visualization/presentation
  • If no, check your data if there is a possible mistake/outliers.
  • If none, ask your colleague to do another analysis to validate your analysis (Feedback)
  • if there is a mistake, go **Clean, Wrangle, Transform

Data Visualization Steps

• Define the data you needed (Raw or Aggregated)
  • For Raw data, geospatial data
  • Aggregated data, correlations
• Choose the appropriate data viz (https://datavizcatalogue.com)
• If needed, decide if you will create a dashboard rather than visualization
• Make a story out of it (Organized all of your analysis, visualization in to a wonderful ) - Reference Discover | Tableau Public
  • Remember: Use the appropriate words also!
  • Then get ready for your presentation deck (Combination of visual story, and oral story) 7 Data Storytelling Tips to Improve Your Presentations - YouTube